I love digital tools that have some analog look or feel to them. (Speaking as an Excalidraw super user.) Going to have to find an excuse to use this beauty!

How to multiply your shipping cadence while using agents responsibly. Matthew Binshtok on the Vercel blog:

There is a fundamental difference between relying on AI and leveraging it.

• Relying means assuming that if the agent wrote it and the tests pass, it’s ready to ship. The author never builds a mental model of the change. The result is massive PRs full of hidden assumptions that are impossible to review because neither the author nor the reviewer has a clear picture of what the code actually does.
• Leveraging means using agents to iterate quickly while maintaining complete ownership of the output. You know exactly how the code behaves under load. You understand the associated risks. You’re comfortable owning them.

I’ve seen a lot of strong opinions about disclosing whether code in a PR was written by hand or generated by AI. I don’t really care. The author owns the code in the first place. The author and reviewer have a shared responsibility for what happens on production.

Putting your name on a pull request means “I have read this and I understand what it does.” If you have to re-read your own PR to explain how it might impact production, the engineering process has failed.

The litmus test is simple: would you be comfortable owning a production incident tied to this pull request?

It's a day that ends in "y", which means The Verge is at it again: covering the regulatory tussle over mobile ecosystems without mentioning browsers, the web, or PWAs.

This long-established pattern is as frustrating as it is predictable. Silicon Valley's journalists fail readers by neglecting to connect the dots between the forces that kept desktop OSes open over the past two decades and the suppressive tactics today's mobile duopolists¹ deploy to prevent similar outcomes. And this despite connections handed to them on a platter by OWA and others. In 2026, the choice to cover app store regulation exclusively through the lens of steering and native app alt stores is a yawning blind spot.

Apple (in particular) is going to extreme lengths to prevent the emergence of true browser choice as a way to obstruct the app store's most credible disruptor. Capable browsers, and the PWAs they support, hold the power to grow an ecosystem of applications that no gatekeeper can own or tax, based on standardised APIs that resist enclosure. But few outlets are connecting these dots for readers.

Cupertino is terrified of the web's potential, but you'll never read about it on the pages of Wired, MIT Tech Review, The Verge, or 404. Their readers are also entirely uninformed about how amenable regulators have been to the gatekeepers' stalling and misdirection. Only The Register and TNS understand what time it is, covering the essential points regularly:

• Apple's succeeding in delaying real browser choice through underhanded and unlawful tactics
• Regulators are not enforcing the laws effectively in these areas
• Users and local businesses all lose as a result

These regulatory and journalistic failures are not victimless, circumscribing what technology can be for us all, both positive and negative.

Open Societies Necessitate Interoperability

In 2026, it should go without saying that billionaires at the helm of BigTech firms cannot be trusted to do anything but cower and pirouette for Mad King Trump. The "good ones" have been proven feckless when presented with opportunities to defend democracy and moral decency. The rest are either actively fascist, fascism-curious, fascist-enabling, or so engrossed in tax dodging they haven't clocked the rule of law succumbing to attacks by their peers. Not that there's much hope they'd stand up for it; they've got theirs, Jack.

These are the masters of the universe demanding control over what happens on our most personal computers.

In an era when they still claimed to be changing the world for the better, the lies told to justify an invasive, unsafe ecosystem of native apps went down more easily. The duopolists provided shockingly deep hooks into our lives for native app developers while shielding them from adversarial interoperability. In exchange, they demanded control over distribution and taxed the pants off the gambli... — sorry, "casual games" — and privacy degrading social media apps that powered the growth of the native app ecosystem.

Accomplishing this feat required defeating threats from open technologies, because open systems allow users to assert their interests in the face of abusive and predatory software. If users had a vote where useful software could be acquired, or could use browsers for more of their needs, it wouldn't be possible to tax developers at such a high rate. The anti-privacy, anti-user agenda that Apple and Google draw excess profits from cannot survive open technology. This is reason enough for democracies to demand portable, standards-based software.

Democratic societies have wider interests too, and we see clearly that trusting Apple and Google with as much market power as they demand is a threat to those goals. The shocking ethical derelictions of the app stores provide yet more reasons to demand regulators legalise open technology.

Software and digital services benefit from network effects, creating acute winner-take-all dynamics in markets where countervailing forces are suppressed. Traditionally, those crosswinds for would-be monopolists have included over-the-top applications, Open Source software, adversarial interoperability, and end-user modulation through open platforms (e.g., ad-blocking browser extensions). Over time, these disciplining factors can break the hold of short-term monopolists, returning power to users and forcing abusive proprietors of services to do better by society or risk losing market share as moats dry up from the pressure.

The iOS and Android duopoly have rendered these recourses illegal and impractical, minting a small number of durable winners. The CEOs and founders of these colossuses are the new tech oligarchs, and their actions in recent years provide compelling arguments against both money in politics and the idea that there can be such a thing as a "good" billionaire.

The broligarchs have shown no compunction towards subverting the state for illegitimate private ends. Having corrupted American politics with torrid rivers of cash, news of their support for anti-democratic, authoritarian regimes around the world no longer surprises. Shocking evidence of this corruption is on display almost daily. Not content to curdle American governance, the broligarchs now demand others desist in enforcing their laws lest they dampen their accelerating profits.

This corrosive effect is only possible thanks to the intense concentrations of wealth enabled by perverted market structures. These twisted markets are themselves the result of decades of neo-liberal anti-antitrust groupthink. Contra Bork, the social contribution of robust antitrust enforcement isn't just lower prices, it's the promotion of stable institutions and honest dealing across a wide swath of society. Samuel Bagg frames the opportunities for antitrust enforcement similarly:

...many practices of constitutionalism are justified on similar grounds: i.e., that first-order constraints on the abuse of power can only be sustained through second-order limits on the concentration of power more generally. Where modern constitutionalism aims primarily to limit the concentration of public power, however, the anti-monopoly tradition follows earlier materialist forms of constitutionalism in viewing the concentration of private power as equally dangerous (Andrias 2015; Fishkin and Forbath 2014; Khan 2018). After all, problems of capture arise not from the existence of public power as such, but from its relationship to private interests.

— Samuel Ely Bagg,
"The Dispersion of Power, Chapter 6.1"

Tim Wu charts a history of antitrust thinking, noting that:

Back in the 1950s and 1960s, the anti-monopoly enforcers were of a different and much tougher mindset. They were closer in disposition to Theodore Roosevelt and tended to see the stakes of anti-monopoly as transcending mere economics and spilling into questions of democracy and political destiny. Influenced by the recent experience of the Second World War, American officials believed that fascism and corporate monopoly were linked. They feared that excessive corporate power would bleed into fascism or contribute to a communist uprising. Famed antitrust enforcer Thurman Arnold blamed German monopolies for helping Hitler rise to power.

— Tim Wu,
"The Age of Extraction, Chapter 2: Platformization"

Like Wu and other Brandeisian thinkers, Bagg finds counterweights to these corrosive forces useful, both structurally and specifically:

Countervailing power is the natural complement to anti-monopoly. Both demands are set in motion by the insight that capture is most often perpetrated by certain hegemonic actors and groups — i.e., those with the most concentrated private power and organizational capacity — and that protecting the public interest thus entails special attention to those specific forces. Where anti-monopoly entails targeting their resources and organizational capacity directly, however, practices of countervailing power aim to support their opponents.

— Samuel Ely Bagg,
"The Dispersion of Power, Chapter 6.2"

The web is such a countervailing power, bringing with it a large ecosystem of developers and publishers that can jump-start mobile contestability. Real browsers and web apps present a potent, two-sided ecosystem competitor. Because browsers feature low switching-costs compared with replacing phones, and provide a standards-based, interoperable app platform, they carry maximum potential to disrupt the market-concentrating effects of app stores with software catalogues tied to specific OSes. These properties pose a credible threat to the smartphone duopoly's extractive, anti-competitive status quo. Indeed, we can expect real browser competition to finally make Apple and Google competitors, returning economic surpluses to users and businesses, rather than the duopolist's shareholders.

A web liberated to compete has already demonstrated the power to align forces that erode gatekeepers' influence through competition between browsers and their engines. Interoperability and a grounding in royalty-free standards pairs with this competitive engine to expand the penumbra of capabilities that enable new entrants without simultaneously convincing all developers to build for a new OS.

Because browsers commoditise OS features, the competition between them — both internal to, and across the existing Android and iOS deployed base — grows a corpus of essential apps that any new entrant needs (the Bridge Strategy). Instead of forcing a new smartphone OS to convince every software house to build new apps for their new APIs (simultaneously), or to brave legal and technical threats as they implement Apple and Google's APIs wholesale, browsers and PWAs shrink the missing software gap tremendously. Should browsers make hay on iOS and Android the way the DMA's drafters hoped, new market entrants only need to port competent browsers (a much smaller lift) and fill in a greatly reduced app gap.

But even if a new competitor doesn't emerge thanks to the web, browsers discipline incumbents in-place by providing competition with their proprietary software ecosystem. When mobile OSes feature sustained browser competition, incumbents help expand the web's carrying capacity or risk marginalisation.

This dynamic played out in the 2000s and 2010s as Internet Explorer lost share to competitors thanks to Microsoft's under-funding. Firefox, Chrome, and Safari worked to expand the power of the web, pushing IE's once 90+% dominance to the curb while dramatically growing the list of tasks users turn to the web for today. Eventually, Microsoft capitulated, switching to Chromium to regain compatibility with a dynamic, growing web that threatened to pass it by.

Proprietary incumbents prioritise user needs over their own extractive agendas when competition from open ecosystems is enabled by low switching costs.² This paves the way for competitors, expanding the power and relevance of low-cost, interoperable technology with each competitive iteration. We have seen the web's wedge work against API enclosure on the desktop, with a large majority of all new desktop software being written first as web applications and only later (and optionally) wrapped in WebViews to deliver needs from a shrinking set of "last mile" capabilities that the web doesn't provide out of the box.

Such is the power of interoperability made manifest; it's far past time for the tech press and regulators to speak clearly about the anti-web game the duopolists are playing to prevent the web from breaking out.

Shared Premises

Several points are plain to anyone engaged in mobile ecosystems. There are many sub-points to quibble about, and various parties talk their own books, but the broad strokes are accepted by all sides:³

• The web provides a portable and (potentially) capable way to deliver software.
• Every contemporary phone with a touch screen has more than enough computing power to run full-fledged browsers that can support capable web apps.
• Yet the smartphones almost every wealthy person carries support only less-capable, lower-quality browsers.
• This suppresses the web's relevance through compounding wealth and quality effects.

A publicly contested (but privately conceded) corollary is that the web is not dominating mobile the way it does desktop because it has not been allowed to compete. From Apple explicitly suppressing competing browsers and breaking critical features OS-wide on the regular, to Google's history of discouraging internal teams from writing mobile web apps and denying competing browsers access to critical PWA features, the fix has been in for 15 years.

Apple and Google's anti-competitive policies keep web apps from threatening the foundations of their App Store revenues by forcing developers to build to the APIs that the incumbents control. This grounding in Apple and Google's proprietary APIs is the root of the duopolist's power. They use it in combination with threats and promotional inducements from their stores to enlist developers in promoting features of the hardware and services they monetise.

Open, interoperable, safe-by-default runtimes with standardised APIs threaten the foundations of this structure by breaking the duopolist's chokehold over software distribution. Browsers, and their core function of abstracting away the proprietary APIs of OS vendors, create a tax-free zone outside the grasp of the OS incumbents' mafioso App Store tactics. To ensure the web cannot threaten the API enclosure today's mobile ecosystems are founded on, browser competition must be suppressed and browsers themselves incapacitated.

That Apple has managed to pull this off is a scandal. But our tech press and commentariat do not cover the web as a true threat to App Stores. Instead, they fixate on alt stores as a mechanism to supplant gatekeeper control. This is a gift-in-kind to the duopolists. By failing to acknowledge the power of the web to disrupt entrenched OSes (as demonstrated by the past 20 years of desktop computing), the press are failing to communicate the stakes of regulatory failure to enable browser competition. This, in turn, excuses woolly regulators as they squander the ripest opportunity for reform in living memory.

Alt Stores, PWAs, and the App Store Endgame

The dream of the alt store is a mirage, or at least a distraction. Windows is able to sustain several successful native app distribution systems in addition to Microsoft's own (e.g., Steam and The Epic Games Store), but these systems are generally tied to specific verticals and the APIs desktop developers target are open. The incumbent's app stores have significantly larger scope and tie together a host of services that require critical mass to disrupt.

The web presents one of the only ecosystems with enough existing use and investment to plausibly re-create a substantial fraction of this ecosystem value without the cold-start problem. Other challengers face a daunting choice: focus on a less expansive vision by diverting into a few vertical app types, or depend heavily on OS services and APIs. Either option reduces contestability and detracts from the potential for a successful, interoperable mobile software ecosystem.

Browsers provide an alternative. Because they synchronise payment instruments, identity, and user preferences over the top the OS, users experience reduced friction when transferring to the web versus a competing native ecosystem. And because most folks already invest a significant fraction of their digital lives with their browser, apps that run on the web are a more credible threat to the status quo than alt stores filled with native apps.

Regulators facing choices about where to invest scarce resources should foreground browser engine choice and web apps because they create mass behind their agenda and bring a powerful, pre-existing ecosystem to bear. The alternative is playing whack-a-mole with gatekeepers over easily divide-and-conquered alt stores.⁴ Once browser-based apps can address the bulk of the market, the lift required to institute fair distribution terms in other verticals will be substantially reduced.

Instead of promulgating proprietary APIs, the web provides an answer to regulators' central problem: how to bootstrap an ecosystem of interoperable software?

Capable browsers are the proven answer. Unleashed by aggressive policy intervention a quarter century ago, the web has become an app store.. As the D.C. Circuit Court predicted:

...were middleware to succeed, it would erode the applications barrier to entry. Because applications written for multiple operating systems could run on any operating system on which the middleware product was present with little, if any, porting, the operating system market would become competitive. Id. p p 29, 72.

But as the District Court found, middleware will not expose a sufficient number of APIs to erode the applications barrier to entry in the foreseeable future.

— US D.C. Circuit Court, Per Curiam,
"U.S. v. Microsoft Corp., 253 F.3d 34 (D.C. Cir. 2001)"

What the court failed to understand in 2001 was that Microsoft had already made the browser a complete platform. Even after Redmond disbanded the IE team after the 6.0 release, it would go on to midwife applications like Google Maps, GMail, and Google Docs; in its zeal to see off Netscape, Microsoft sewed the seeds of Win32's irrelevance. Firefox, Chrome, and (yes) Safari built on the momentum, unlocking adjacent possibilities that brought new categories into the web at a rapid clip. Today, design, CAD, software development, and scientific computing regularly happen in-browser and more than 70% of desktop "jobs to be done" are now handled on the web.

PWAs complete the story for the mobile era, unlocking participation in the tap-and-swipe ecosystem today's duopolists are working enclose through app stores. Just like a quarter century ago, the web is already the answer; we just need to clock it as such and legalise powerful browsers.

So-called alt stores are a less compelling solution because they hold less power to create competition with low switching costs. Instead of providing developers with standardised APIs, they continue to weave the duopolist's proprietary technologies and services into each app. This reduces app portability, increases developer costs, and implicitly helps to build a moat against new OS entrants. A hat-trick of lock-in.

Instead of settling for a world in which alternatively distributed apps must attempt to re-create the technical foundations of interoperability for themselves, regulators can win both interoperability and momentum for their efforts by unblocking real browser choice and PWAs enabled by it. Browser engine projects feature large teams whose entire job is to intermediate between OS-specific APIs and web apps, enabling developers to write to a single interoperable platform out of the box.

The opportunity to harness both a large ecosystem of developers used to building for an interoperable platform, and teams of C++/Rust engineers accustomed to competing to provide improvements to that interoperable platform should have rocketed the web to the forefront of every regulator's thinking. The new powers that the EU's DMA, the UK's DMCCA, and Japan's MSCA provide in theory need concrete enforcement actions to bring the duopolists to heel, and the web's advantages in efficiency of regulatory intervention, reduction in security risk, and pre-existing cross-OS interoperability are unparalleled.

Regulators Are Missing The Ripest Opportunity

It pains me deeply to acknowledge that the institutions handed new powers by the DMA, the DMCCA, and the MSCA are universally failing to prioritise the transformative power of browsers as they work to enforce the provisions of these laws. Instead of focusing fire on the lowest-risk, highest leverage intervention, regulators are being side-tracked and talked down from pursuing aggressive browser choice enforcement.

Consider Apple's opening offer to the EU. The DMA only allows gatekeepers to impose necessary and proportional security restrictions on competing browsers and their engines.

Regardless, Cupertino larded up its package of DMA deliverables with terms for building browsers that include clearly unlawful contractual obligations (pdf) having nothing to do with security. These restrictions levy costly and unworkable demands on the architecture of competitors' products. Alternative browser engines have long featured better security, performance, extensibility, and even privacy than Safari. Should they not be superior, the DMA provides a straightforward solution: competition, not legal arm-twisting.

In case procedural hurdles weren't enough disincentive, Apple also hobbled its browser SDK, omitting support for obviously in-scope features including the ability to handle in-app links, support web apps on the home screen, and disambiguate push notifications. It has built and shipped all of this for Safari, in some cases relatively recently, so it wasn't a case of forgetfulness. Brushed back from attempting to murder web apps on the eve of DMA enforcement, Apple simply chose to duck complying with the law by dressing up violation in umbrage and legalese.

Cupertino's design carefully omits the capabilities that would unlock direct competition with the app store, safe in the knowledge that so long as Safari is the only browser with those features, it can keep them hobbled and broken. This blatantly attempts to subvert the DMA's plain meaning and legislative intent, and dovetails with extralegal attempts to apply maximum pressure on the EC to accept less than half a loaf. Recall that Apple initially argued that it didn't have time to comply given a mere two-year run-up. It was enough time for all of Apple's competitors, but somehow not for the self-proclaimed titans of software from Cupertino.

That was two years ago.

Has the offer improved regarding these strategic omissions? Not one iota.

Indeed, Apple is now lying to our faces; from its latest DMA compliance report (emphasis added):

(45) For the purposes of allowing alternative web browser engines to interoperate with iOS and iPadOS, Apple makes a suite of functionalities available so that browser apps and apps providing in-app browsing using alternative browser engines can be installed and run on iOS and iPadOS in the EU. Apple provides authorized developers access to technologies within the system that enable critical functionality and help developers leverage high-performance modern browser engines. These technologies include Just-in-Time compilation (which provides performance benefits by helping browsers to quickly and efficiently render JavaScript content), multiprocess support, and more. Third-party browser apps using third-party web browser engines can support and run web apps.

— Apple, March 7, 2026 (pdf),
"DMA Compliance Report Summary, Page 201, Section V."

This misleading text suggests that Apple has provided APIs to competitors to support these core use-cases when using their own engines; it manifestly has not.

Cupertino is Running Circles Around Brussels, Tokyo, and London

Faced with blatant, strategic flouting of these new regulations, one might assume regulators with a sense of their own prerogatives and legislator's intent might respond with overwhelming force.

Instead, the DMA team at the EU have moved haltingly as Apple plays for time through processes the EC forced Cupertino to create after previous bouts of transparent gamesmanship (pdf). Clear, unambiguously in-scope requests from competitors languish while the Commission settles for tweaks that cannot change the balance of power. Far from fulfilling the law's legislative intent, contemporary DMA enforcement appears unable to prioritize action that will unleash the tools of revolution from action on necessary-but-insufficient nips and tucks.

Rather than treating Apple's belligerent, maximalist strategy as a joined-up attack aimed to deny oxygen to the most effective competing ecosystem, the DMA's enforcement group has ratified Apple's redrawn boundaries through studious inaction. Like Sykes and Picot with a ruler, Apple has drawn lines that favour its interests above all else, and authorities that could countermand its declarations are notable only in their absence. The perspective from Brussels seems to be that these questions are now simply competing claims by equally mucky mud wrestlers, rather than galling non-compliance and an affront to the entire point of the DMA.

Rather than bearing down on structural issues, regulators appear immersed in dense debates with the scofflaw's counsel over inconsequential details. Even in the areas that it is levying fines, the EC is not challenging the structure of Apple's power. This is tantamount to settling for minor tweaks on the status quo. If the DMA was designed to redress the problems it enumerates, this is failure by the EC's own yardstick. We're entering year three of DMA enforcement, and meaningful change is not obviously closer. Why not?

To hear the regulator's apologists tell it, they're concerned with litigation. Litigation! As though Apple wouldn't sue regardless.

No amount of punch-pulling or case buttressing will possibly prevent Cupertino from rolling out its shockingly large army of lawyers and worse. Apple already spends more on lawyers every year than the fines the EU is proposing!

This is the same company that used lawyerly nonsense to hold up the UK CMA's investigation for nine months, implausibly arguing the regulator hadn't brought the boom down fast enough on its own bad behaviour, and therefore couldn't do so later. It was never a serious argument, but when dressed up in enough £1000/hr legalese, Apple's calculated attempt to create delay paid massive dividends. Nine months of regulatory delay is worth many billions to Cupertino.

Cupertino is conducting a maximalist campaign against the very concept of the rule of law when its profits are at risk, and having paid off a deeply corrupt US administration, the intensity of that campaign is only increasing. All of this should create a deep, abiding suspicion of Cupertino's motives, removing the benefit of the doubt wherever it claims contestability is a reality on iOS. No self-respecting regulator can take anything Apple's lawyers say at face value in 2026. But there's only a price for duplicity when those with authority recognise the pattern and apply suspicion, rather than deference.

Regulators (not to mention the press) are well within their rights to take on board Apple's misrepresentations, astro-turfing, and serial non-compliance when calculating how aggressively to act. Instead, the EU's DMA team, the UK's CMA, and Japan's regulators are falling for the same shenanigans over and over; treating each encounter as a blank slate, happy to pretend a disquieting history of belligerent non-compliance is immaterial to arguments offered today.

This is both a tactical error and a strategic blunder.

At a tactical level, extending deference to each new pile of misdirections allows Cupertino to play for time; when you're the incumbent, delay is winning, and regulators that do not demand prompt changes weaken their own future hand. Strategically, failure to demand changes right bloody now in the face of material non-compliance denies the regulator the benefit of even a weak fleet in being effect. And declining to demand immediate change in areas most likely to create the deepest, most generative effects grants Apple ongoing regulatory relief in the form of reduced structural competition.

Of course it harms EU businesses that they cannot target a platform across mobile OSes, and that EU businesses instead have to bear high ongoing development and servicing costs to participate in proprietary app stores. It is obviously negligence on the part of newly empowered regulators to allow Apple's gerrymandering attempts. Cupertino has worked to draw constrictive borders around browsers for the last several years; demarcations that just happen to exclude capabilities that would make mobile computing truly contestable.

Browsers Are App Stores

Since regulators and the tech press are not connecting the dots from first principles, desktop computing analogues, legal precedent, or the lengths Cupertino has gone to prevent the web from breaking out, it must be repeated:

Browsers are app stores.

If want effective alternatives to the gatekeeper's app shops, nothing will be more effective than unleashing real competition from the web. Browsers are stores filled with a priori interoperable software and with low friction installation thanks to PWAs! And they feature incredible security! Protections so strong that you don't even have to hang a "beware of dog" sign out front.

Should we collectively accept this premise and expand our understanding of software contestation to include the web as a viable alternative in the app store story, what would that mean? The duopolists are working overtime to silo the discussion to their preferred proprietary technologies, why? Is that worthy of investigation and discussion?

For regulators, this perspective is a treasure map. Enabling true browser competition unlocks both internal and cross-OS competitive forces, unleashing firms with an interest in interoperability on both sides of the browser market. But for regulators to get credit for converting on this incredible opportunity, their constituents will need to hear about it. And on that front, we've got a lot of work to do.

Our Tech Press is Missing the Story, Too

We get plenty of TopGear-but-for-phones and "CEO said a thing" coverage, but very little coverage of the way access to APIs shapes our digital lives. Most outlets still treat access like it's a real beat, depriving them of the distance necessary to contextualise the duopolists' actions in a frame their PR staff dislike. The closer a publication gets to the mainstream, the worse the distortion seems to get.

Few tech outlets, including irreverent youtubers, seem to recognise that the web is the single biggest threat to mobile app stores, opting instead to cover minor news of alt stores tremors like The Big One. This misleads readers about the stakes and distorts the incentives of regulators.

Reporting that borrows the narrative frame the duopolists prefer enlists otherwise excellent outlets as allies of the billionaire class. Too many reporters uncritically accept that app stores are how we must think about mobile software, regurgitating this narrative through stories filed in a browser on their laptops. Apple, Google, and the handful of "winners" the app stores bless profit massively from centralisation. Reporters and editors further the goals of the broligarchs by failing to reframe mobile competition in a more historically grounded light.

Perhaps regulators will find their way without the help of a thoughtful, tenacious tech press; stranger things have happened. But judging by The Verge, Wired, and the FT, they're going to need all the luck in the world.

FOOTNOTES

1. Understood through the lens of price segregation, we can see that there is no real competition between Android and iOS. For more than a decade, the mobile market has presented a pantomime of competition, wherein all wealthy users buy iPhones and everyone else buys Androids, while OEMs not named Apple continue to send low-volume "halo" Android devices to reviewers to keep up the appearance that they are somehow a challenge to Apple's dominance of the high-margin premium and ultra-premium market segments.

   In reality, there is no such competition. Apple and Google preside over co-monopolies, separated like oil and water by point-of-sale device pricing, and both understand the other will not challenge the other's segmented market position.

   For Apple, this assurance comes from the financial model for its entire enterprise. Apple cannot (and does not) enter lines of business where it cannot guarantee at least 30% net margin. This is a stable property of Apple's operating behaviour that market analysts now rely on. Any deviation from trend would not enhance Apple's market value, but would instead detracting from traders' ability to understand Cupertino’s business prospects.

   Google, for its part, fears Apple could release a low-margin iOS device, while yearning to release an "iPhone Killer" of its own. For similarly structural reasons, it never will.

   The Pixel brand cannot succeed in a similar way because the firm will never allocate enough capital to make serious attempt to vertically integrate the way Cupertino does. This leaves it — along with the rest of Android ODMs — without sufficient leverage and integration discipline to produce devices with similar quality, longevity, and post-sale monetisation upside. Other Android OEMs face similar challenges, but with even smaller war chests and worse prospects for vertical integration. As long as Qualcomm has everyone over the barrel (and they do), it will never be possible for an Android OEM to break the structure of the duopoly. Android is therefore destined to define low-margin devices. Knowing this drags all Android participants away from post-sale quality improvements that might allow the ecosystem to aspire to better. This sucks for OEMs and ODMs, but the volume of the business makes the structure stable. ⇐

2. It's no coincidence that Apple's basic stance towards browser competition is that you can have a better mobile browser, you just need to buy a different phone, suffer the loss of all of your Apple-ecosystem services and data, and worry about the availability of important apps.

   High and uncertain switching costs are core to Cupertino's suppression of browser competition, and it's bewildering that neither the tech press nor regulators seem to have clocked the gambit for what it is. ⇐

3. There are, of course, multiple ways to speak about these undisputed facts, and native app partisans are particularly practised at their variant of three-card monty.

   When pressed, some begin reciting a list of missing PWA capabilities. This is offered as a reason to prefer native apps, and a reason why web apps are "not serious".

   If handed evidence that open, interoperable versions of these features are not available because of Apple's suppression, they will claim that it is not safe for browsers to offer them. This is often couched as evidence of Cupertino's far-reaching wisdom.

   Evidence that Apple's alternate — authoritarian control over APIs and software distribution — has not been particularly safe is usually dismissed out of hand. Point out that most energy invested in APIs over the past decade has worked to attenuate their power to the position browsers begin the bidding, and native app partisans reframe this work as proof of the necessity of gatekeepers, rather than evidence of their frequent fallibility.

   There are also generally digressions into privacy, where native app fans seem unable to understand that it was the duopolists that provided unfettered access to shocking amounts of invasive data to native apps by default. Despite performative curbs, this situation largely persists, facilitated by the gatekeepers. The idea that we should trust them now, when they are the same unrepentant players that took shockingly little care when handing unrestricted APIs to tech's worst actors, is generally parried with "but Facebook bad!". Kayfabe is treated as serious drama. The deeper complicity is not engaged, and this line of inquiry is generally fruitless. Many wear the brand of phone in their pocket like a jersey for their favourite team, not a complex business with many incentives and cross-pressures.

   These counterparties are also generally unfamiliar with the ads ecosystem and how the native duopolists created a "snoopers gap" that generated pressure for ad-funded apps to push users away from more privacy-respective, adblock-enabling browsers, and towards less mediated and transparent native app alternatives. The reality that those prompts were not the result of an intrinsic positive technology difference is not seriously entertained.

   The web has a long track record of protecting users more effectively on both privacy and security than native API monocultures, and providing extension points to allow users to further shield themselves without waiting on beneficent App Store proprietors. Should that reality make a dent, the next conversational shuffle turns to browsers not already having won the field. This is couched either as evidence of some ineffable technical inferiority on the web's part or a claim for why it wouldn't be possible for a transition to start now; never mind that both Windows and macOS made similar turns on a similar time scale. It is never accepted as evidence of active suppression of the web by the gatekeepers.

   Which brings us back to the start of the game.

   Having played many pointless rounds with folks who, on the surface, seemed open to discussing the issue, I can only recommend checking to see what sort of phone your counterparty carries before wading in. iOS partisans are particularly attached to the idea that Apple knows best. It is often counterproductive to engage, no matter how loudly they claim to support the web.

   One of the benefits of being locked in the privilege bubble, after all, is freedom from introspection. Wealth's blanket soothes the anxieties of the status-seeking mind, and supporting the nonsensical positions by those with higher status is imagined to be a harmless bit of in-group signalling. ⇐

4. It didn't take a genius to predict that a successful alt store would prod the duopolists to make life inhospitable for developers trying to straddle open ecosystems where they are available and gatekeeper options where they aren't. The unpredictable element was that Apple played those cards before a store even got going in earnest. ⇐

Nicholas Carlini, a research scientist at Anthropic, reported at the [un]prompted AI security conference that he used Claude Code to find multiple remotely exploitable security vulnerabilities in the Linux kernel, including one that sat undiscovered for 23 years.

Nicholas was astonished at how effective Claude Code has been at finding these bugs:

We now have a number of remotely exploitable heap buffer overflows in the Linux kernel.

I have never found one of these in my life before. This is very, very, very hard to do.

Musings from Robin Sloan:

Most olive oil production at medium-or-greater scale depends on machines of this kind [over-the-row olive harvester]; they trundle over trees planted in long rows, almost like continuous hedges, and collect the fruit with vibrating fingers. Machine-harvested olives cost less to buy, and they arrive at the mill in better shape than olives harvested by hand.

The catch: most olives can’t be cultivated in this configuration; the trees don’t thrive so close together. Only a handful of varieties will tolerate it, so those handful have been planted in huge numbers, and the flavor of global olive oil has changed as a result.

AI enables us to do things faster, and sometimes better than we’ve been able to before. But it has its limits. And as we learn those limits, the work we do will shift to avoid them.

In a different section, the article dives into the limitations of the physical world.

The project cut across several different magic circles — Ruby code, quasi-governmental APIs, the rules and standards of the postal system — but/and it also broke out into the physical world of paper, printers, and post offices. The project required manipulations including but not limited to: folding, peeling, sticking … gnarly!!

It’s possible that an AI coding agent could have helped me with #1 above, and of course it could have advised me on the rest. But it’s impossible to imagine the AI agent handling #2-5 autonomously; it would require such a Rube Goldberg tangle of support that “autonomously” would no longer apply.

In our programming world, AI’s impact looks limitless. But once you drift outside the boundaries of software, it’s put into a different perspective.

If indeed AI automation does not flood fill the physical world, it will be because the humble paper jam stood in its way.